This page reports my current thinking on the topics that I work on. For the list of published works, please consult the papers and talks pages.
I am interested in morphosyntactic conditions of allomorphy
- Locality of allomorphy. I believe that the trigger and the target of allomorphy must be the structurally adjacent: they must be sisters in the syntactic structure. My current thinking is that the counterexamples should be accounted for by positing that morpheme orders can present structural ambiguity.
- Cumulative exponence. I believe that morphosyntactic heads can correspond to a single exponent iff they form a ‘‘chain’’ of being pairwise structurally adjacent. This differs both from theories that limit cumulative exponence to linear or head-complement sequences. I am currently looking into cumulative exponence of pronominal arguments with tense-aspect-mood features as evidence in favor of my view.
I am interested in linearization of syntactic structure
Head movement and affix hopping. I believe that a multidominant theory of head movement can be extended to cases of so-called affix hopping (also known as Lowering) if formation of complex words is sometimes driven by selectional requirements of the root-categorizer combination. While this requires a partial dissociation of word formation and phrase formation, I think it is required by recent work on cross-categorial derivations (the small syntax approach, which I believe to be well motivated).
Affix order and word order. As mentioned above, commonly held view is that word structure and affix structure are not qualitatively distinct. This raises questions regarding affix order and word order: whether head-complement orders are the same inside words and between words, whether generalizations like FOFC hold word-internally, and whether phrasal movement can account for problematic affix orders.
Interactions with ellipsis. Constraints on syntactic movement have been argued to follow if linearization applies cyclically, with incoherent linearizations resulting in ungrammaticality. If so, ellipsis is predicted to repair violations of such restrictions. Together with Ivan Kalyakin (ILS RAS), I look into Russian left branch extraction which presents a case of repair by ellipsis but necessitates a more intricate theory of ellipsis-linearization interaction because repair by ellipsis is not freely available.
I am interested in phonological conditions of allomorphy:
- Utility of underlying representations. I believe that abstract underlying representations provide a better model for complex morpho-phonological phenomena than alternatives like transderivational comparison of outputs or direct reference to syntactic structure by phonological rules. My research goal is not only to provide abstract UR analyses but to also explore their learnability and compatibility with experimental results on productivity.
- Phonological conditioning of suppletion. I believe that we currently lack a satisfactory theory of cases when a suppletive alternation refers to prosodic structure: phonological optimization approaches are empirically inadequate, while subcategorization frame approaches are theoretically unsatisfactory due to their lack of predictive power and a problematic conception of contexts of insertion rules vis-a-vis natural classes.
I am interested in exhausitifcation
- Polarity items. I am working on an exhaustification-based proposal for Russian positive polarity disjunction i which relies on its competition with conjunction i: in my FASL 34 talk, I show that the disjunction i is able to take scope under negation in cases where the conjunctive alternative is either contradictory or pragmatically irrelevant.
- Triviality. Exhaustification is unable to rescue trivial sentences from ungrammaticality. Why?
- Ellipsis. It seems to mess up alternative-sensitive operators: only cannot associated inside an ellipsis site, Exh qua a syntactic entity is not detected by ellipsis, and work on sluicing with else-correlates suggests that questions with clausal ellipsis are not as exhaustive as those without. Again, why?